title:: Patient-specific Modeling and Quantification of the Heart Valves from Multimodal Cardiac Images (Doctoral Dissertation, TUM)
publisher:: Technical University of Munich
people:: Razvan Ionasec, Nassir Navab, Nicholas Ayache, Dorin Comaniciu
organization:: Technical University of Munich, Siemens Corporate Research, INRIA
domain:: Medical Imaging, Machine Learning, Computational Modeling, Cardiology
link:: https://mediatum.ub.tum.de/981719
citations:: Cited 2 (OpenAlex) · 3 (Semantic Scholar) — as of 14 October 2026
Summary
The doctoral thesis that frames the whole decade of work. Three aspects: physiological heart valve modelling, algorithms for patient-specific parameter estimation from multimodal images, and advanced applications for the clinical management of heart valve disease.
It introduces a physiological model of the complete valvular apparatus capturing a large spectrum of morphologic, dynamic and pathologic variation, and a robust learning-based framework to estimate the patient-specific model parameters from transesophageal echocardiography, cardiac CT and cardiac MRI. The model-driven approach enables quantitative valvular analysis, shape-based diagnosis and patient selection, and computational decision support for valve replacement and repair. Against the gold standard - manual processing of 2D images - the automatic 4D model-based approach is faster, more precise, reproducible and more complete.
Supervised by Prof. Nassir Navab (TUM CAMP), Prof. Nicholas Ayache (INRIA Sophia Antipolis) and Dr. Dorin Comaniciu (Siemens Corporate Research, Princeton).
Data Points
- Submitted 30 June 2010; accepted by the TUM Faculty of Informatics 28 September 2010
- Supervisors: Prof. Nassir Navab (TUM CAMP), Prof. Nicholas Ayache (INRIA Sophia Antipolis), Dr. Dorin Comaniciu (Siemens Corporate Research, Princeton)
- Validated through quantitative and clinical experiments involving over 476 patients treated in leading medical centres worldwide
- Physiological model of the complete valvular apparatus across morphologic, dynamic and pathologic variation
- Learning-based patient-specific parameter estimation from TEE, cardiac CT and cardiac MRI
- Applications: quantitative valvular analysis, shape-based diagnosis and patient selection, computational decision support for valve replacement and repair
- Targets percutaneous and minimally invasive valve interventions via automated patient selection and virtual valve implantation
- Full text: mediatum.ub.tum.de/download/981719/981719.pdf
Reference
Technical University of Munich, Faculty of Informatics Doctoral dissertation, 2010. mediaTUM 981719
Submitted 30 June 2010; accepted by the TUM Faculty of Informatics 28 September 2010
Part of the complete record in Scientific Publications - Peer-Reviewed Research Record.
